Blueprint: How I Built the AI Lead Finder & Marketing Audit Agent

Check the AI Bot Here

If you haven't tried my AI tool yet, Click here and Try it out NOW for FREE!

1. Objective

✅ Build an AI-powered tool that:

• Finds local businesses in any location & category.

• Performs a quick marketing audit: website speed, UX, SEO, social media presence.

• Generates smart outreach hooks & suggestions.

✅ Make it fast, user-friendly, and accessible via a simple web app (no coding needed for end users).

2. Technology Stack

Layer	Tools Used	Purpose
Frontend + UI	Streamlit	To build a lightweight, Python-based interactive app
AI / NLP	OpenAI GPT-3.5 via openai lib	To analyze business websites and generate insights
Data Source	SerpAPI for Google Maps	To fetch local business listings dynamically
Python Libraries	requests, pandas, json	For API calls, data processing, and structuring tables

3. Core Components

a) Local Business Data via Google Maps

• API used: SerpAPI

• Endpoint: google_maps engine

• Data fetched:

  • Company Name

  • Website URL

  • Type / Category

  • Ratings & Reviews

This allowed the agent to dynamically pull businesses based on:

b) Marketing Analysis with GPT

• API used: OpenAI (via the openai Python package)

• Model: gpt-3.5-turbo

  
  

Prompted GPT to:

• Act as a digital marketing consultant.

• Review the business type, name, and website.

• Return a JSON with fields like:

  • insight: overall marketing note

  • hook: outreach line

  • speed: guess on site speed

  • theme: UI / UX theme suggestions

  • seo: quick SEO improvements

  • social: social presence guess

    
    

This made it easy to parse into a neat pandas DataFrame.

c) Classification of Ratings

Added a simple helper:

This helped highlight brands that might be underperforming in reputation.

4. Building the UI with Streamlit

Used Streamlit because it:

✅ Requires minimal setup.

✅ Allows python + markdown + HTML styling.

✅ Easy to deploy on Streamlit Cloud.

Added custom styling with:

• st.markdown and embedded <style> to change background colors, buttons, and text colors.

• Download CSV button to let users save the leads.

5. How it all connects

6. Key Features Implemented

✅ Enter any business type & location — dynamic results.

✅ One-click Get Leads to process.

✅ AI marketing audit runs instantly.

✅ Download leads as CSV.

✅ Clean, responsive UI with custom theme.

7. Challenges & Solutions

Challenge	How I solved it
Avoiding API key exposure	Stored keys in secrets.toml (Streamlit native way)
Handling bad JSON from GPT	Used try...except for json.loads()
Rate limiting on GPT calls	Added time.sleep(1) between calls
Styling Streamlit cleanly	Embedded CSS inside markdown

8. Tools Recap

Tool	Why I chose it
Streamlit	Quick Python-based UI & deploys easily
SerpAPI	Simplifies scraping Google Maps data
OpenAI GPT	Strong text analysis & suggestions
Pandas	Structure data & generate CSVs
Requests	Easy REST API calls

9. How You Can Try or Customize It

✅ Clone my GitHub repo

✅ Put your API keys into .streamlit/secrets.toml

✅ Done!

Need Any Help? Connect with Me!!!